Pages

Jump to bottom

20 comments

1 wrenchwench  Sat, Mar 15, 2014 5:00:04pm

Well done.

2 RealityBasedSteve  Sat, Mar 15, 2014 6:44:36pm

I like this article 100% +/- 3.5%. I do agree that there are a LOT of people who don’t understand statistics at all, and why they can be extremely valuable or absolutely worthless depending on how they were gathered, interpreted and presented.

Looking forward to installment 2

RBS

3 Political Atheist  Sat, Mar 15, 2014 6:57:21pm

Obdi, by population do you mean people only or also products as in product quality controls?

4 b_sharp  Sat, Mar 15, 2014 7:00:37pm

re: #3 Political Atheist

Obdi, by population do you mean people only or also products as in product quality controls?

Anything you want to get information from.

You can count turtles with round shells, with squarish shells, with oval shells if that is the info you are interested in.

They don’t have to be able to answer questions.

5 Fairly Sure I'm Still Obdicut  Sat, Mar 15, 2014 7:01:04pm

re: #3 Political Atheist

Obdi, by population do you mean people only or also products as in product quality controls?

Also products. Anything. I’ll make that clear, good point.

6 b_sharp  Sat, Mar 15, 2014 7:03:49pm

A population is the total number of things you want to study. Since it is impossible in most cases to count/analyse the entire universe a sample is taken and analyzed.

7 Political Atheist  Sat, Mar 15, 2014 7:04:05pm

re: #5 Fairly Sure I’m Still Obdicut

We do some delicate and detailed sampling in the jewelry mfg biz. I look forward to picking up any tips that might apply. Never stop learning right?

8 Kruk  Sat, Mar 15, 2014 7:04:35pm

I regret I only have one upding to give this page. I spend a lot of time helping design audits, surveys and small scale interventional studies at the hospital I work at. I find even health professionals (who would have taken statistics courses as part of their degrees) struggle with this stuff.

9 b_sharp  Sat, Mar 15, 2014 7:05:23pm

BTW, Obdi, I am really glad you are doing this.

10 Fairly Sure I'm Still Obdicut  Sat, Mar 15, 2014 7:08:16pm

re: #7 Political Atheist

We do some delicate and detailed sampling in the jewelry mfg biz. I look forward to picking up any tips that might apply. Never stop learning right?

Well, the tip I’d have for you then would be to make sure that if you are using any systematic sampling (every nth item is sampled) that you combine it with random sampling to make sure that some pattern isn’t slipping by you.

11 Political Atheist  Sat, Mar 15, 2014 7:35:57pm

re: #10 Fairly Sure I’m Still Obdicut

Our industry has by necessity gotten pretty good at it. My small part is “pin sampling” from a great big melt, maybe a kilo or two of scrap gold. What we do as a refinery is a pin sample. That is a glass vacuum tube that sucks up molten metal from the big crucible in the furnace. That big melt will become a few bars that will be nth checked (via a small drilling from the bars) but should in aggregate match the pin sample. In the end we have the cross check of the actual refining result. Much cross checking occurs.

Gold is heavy so it settles as it freezes. Refining margins are under a percent so accuracy is a must. Often the melt is paid for based on assay not end result.

From time to time we can mess up but the money drives the careful nature of the sampling and cross checking.

12 Fairly Sure I'm Still Obdicut  Sat, Mar 15, 2014 8:14:30pm

re: #11 Political Atheist

As long as the pin sample is coming from a random part of the crucible, or the mixture in the melt is uniform, that’s a simple random sample.

13 Political Atheist  Sat, Mar 15, 2014 8:29:46pm

re: #12 Fairly Sure I’m Still Obdicut

Random indeed. So much so it appears the method tests the accuracy of the assay rather than the uniformity of the blend.

There is a technical detail, the induction current and convection force a very uniform blending of the metals. Think extremely well shaken milk and chocolate.

I look forward to the rest of your series. Going to cover confidence interval?

14 b_sharp  Sat, Mar 15, 2014 8:36:50pm

Not to step on your toes Obdi.

The aim of any sampling strategy is to represent the population (universe) as accurately as is possible without introducing a bias that can skew the sample away from that ideal.

As Obdi did in the response to Political Atheist, sampling strategies are tailored to give that accuracy in specific cases.

15 b_sharp  Sat, Mar 15, 2014 8:37:43pm

re: #13 Political Atheist

Random indeed. So much so it appears the method tests the accuracy of the assay rather than the uniformity of the blend.

There is a technical detail, the induction current and convection force a very uniform blending of the metals. Think extremely well shaken milk and chocolate.

I look forward to the rest of your series. Going to cover confidence interval?

Has to.

16 Fairly Sure I'm Still Obdicut  Sat, Mar 15, 2014 8:52:46pm

re: #13 Political Atheist

I look forward to the rest of your series. Going to cover confidence interval?

Yep, along with standard deviation.

17 Floral Giraffe  Sat, Mar 15, 2014 11:59:53pm

OK, this is awesome.
And this

(The No-Gown-Clown-After-Sundown-in-Our-Town Amendment

cracked me up!

18 FemNaziBitch  Sun, Mar 16, 2014 5:12:57am
This is the main reasons why social science is a ‘soft’ science. It is not just that things are studied that are subjective, it’s also that with even the objective things, obtaining a perfectly random sample is difficult.

THIS!!!

Also, what I noticed most from your post, although I’ve never taken a statistics class, is now much it resembles basic Logic -(philosophy school). Conceptually, anyway.

Is Logic the alternative for those of use who are lost in the mathematical calculations?

19 b_sharp  Sun, Mar 16, 2014 10:23:11am

re: #18 FemNaziBitch

THIS!!!

Also, what I noticed most from your post, although I’ve never taken a statistics class, is now much it resembles basic Logic -(philosophy school). Conceptually, anyway.

Is Logic the alternative for those of use who are lost in the mathematical calculations?

Only in that they both can be expressed using the structure of math. The math in stats only gets really sticky when it gets to probabilities.

20 Fairly Sure I'm Still Obdicut  Sun, Mar 16, 2014 12:06:33pm

re: #18 FemNaziBitch

Is Logic the alternative for those of use who are lost in the mathematical calculations?

As b_sharp says, only in some ways. For example, in a normal distribution, 68% of the population is within 1 standard deviation of the mean, 95% are within 2. That’s hard to communicate in pure logic.


This page has been archived.
Comments are closed.

Jump to top

Create a PageThis is the LGF Pages posting bookmarklet. To use it, drag this button to your browser's bookmark bar, and title it 'LGF Pages' (or whatever you like). Then browse to a site you want to post, select some text on the page to use for a quote, click the bookmarklet, and the Pages posting window will appear with the title, text, and any embedded video or audio files already filled in, ready to go.
Or... you can just click this button to open the Pages posting window right away.
Last updated: 2016-01-01 10:29 am PST
LGF User's Guide RSS Feeds Tweet

Help support Little Green Footballs!

Subscribe now for ad-free access!Register and sign in to a free LGF account before subscribing, and your ad-free access will be automatically enabled.

Donate with
PayPal
Square Cash Shop at amazon
as an LGF Associate!
Recent PagesClick to refresh
Who Is, or Was, the Greatest Martial Artist in History?I only got to the rank of 2nd degree black belt in Chinese & American Kenpo, so I am certainly no expert. I earned all my higher belts at American Karate Studios in Northeastern Ohio, and was a competitor on ...
Samuel Vargo
5 days, 19 hours ago
Views: 345 • Comments: 4 • Rating: 2
Tweets: 0 • Share to Facebook
Shares: 0
Comments: 0
: 0
The Way We Regulate Self-Driving Cars Is Broken—here’s How to Fix It The key issue is this: the current system is built around an assumption that cars will be purchased and owned by customers. But the pioneers of the driverless world—including Waymo, Cruise, and Uber—are not planning to sell cars to ...
Thanos
5 days, 22 hours ago
Views: 333 • Comments: 0 • Rating: 1
Tweets: 0 • Share to Facebook
Shares: 0
Comments: 0
: 0
Android’s Trust Problem Isn’t Getting Better Android security is largely failing due to a smorgasbord of issues and lies. Published today, a two-year study of Android security updates has revealed a distressing gap between the software patches Android companies claim to have on their devices ...
Thanos
1 week ago
Views: 591 • Comments: 3 • Rating: 2
Tweets: 1 • Share to Facebook
Shares: 0
Comments: 0
: 0
Tour of the Moon in 4K Take a virtual tour of the Moon in all-new 4K resolution, thanks to data provided by NASA's Lunar Reconnaissance Orbiter spacecraft. As the visualization moves around the near side, far side, north and south poles, we highlight interesting features, ...
Thanos
1 week ago
Views: 562 • Comments: 0 • Rating: 2
Tweets: 0 • Share to Facebook
Shares: 0
Comments: 0
: 0
Vermont Governor Signs Sweeping Gun Control Measures Vermont Gov. Phil Scott on Wednesday signed sweeping gun control measures -- including limits on the size of magazines -- that the Legislature passed last month after contentious debate. The measures:-- Raise the minimum age for gun buyers to ...
Thanos
1 week, 1 day ago
Views: 649 • Comments: 1 • Rating: 1
Tweets: 1 • Share to Facebook
Shares: 0
Comments: 0
: 0
Roseanne Barr Is Tweeting About QAnon, a New Pizzagate-Style Conspiracy Theory Barr's been full metal Trump-loving wingnut in a social liberal suit for a long long time. This is why I would never watch her show, and why I won't be spending money with advertisers on her show once someone ...
Thanos
2 weeks, 6 days ago
Views: 1,940 • Comments: 0 • Rating: 2
Tweets: 1 • Share to Facebook
Shares: 0
Comments: 0
: 0